Summary of Work: Progression of cancer to the metastatic state involves multiple cellular and genetic changes. Using somatic cell hybridization, we demonstrated that acquisition of metastatic ability requires both the loss of metastasis suppressor functions and the activation of oncogenes. We cloned a gene, termed KAI1, that can suppress the metastatic ability of a rat prostatic cancer cell line. We are investigating the ability of this gene to affect the metastasis of human prostate, breast, ovarian, and colon cell lines. KAI1 encodes a 267 amino acid protein with four hydrophobic transmembrane domains and a large extracellular domain with three potential N-glycosylation sites. We have found that KAI1 is highly expressed in normal prostate and breast epithelial cells but is dramatically lower in cancer cell lines derived from metastatic tumors. We have also found reduced or altered protein expression in human bladder, endometrial, lung, colon, ovarian, and melanoma cancer cell lines. We are currently testing new antibody reagents that will allow study of KAI1 expression in archived tissue sections. We have collected matched pairs of normal prostate and tumors, as well as primary and lymph node metastases from breast cancer patients to investigate whether KAI1 expression can be used to predict the metastatic ability of primary cancers. We have shown that mutations can not account for the observed loss of KAI1 expression; we are currently investigating other hypotheses that would cause KAI1 loss of function. We are also identifying KAI1 binding partners in order to determine how they may regulate KAI1 function. We are examining known, candidate proteins, that may interact with KAI1 at the cell membrane, but are also attempting to identify novel partners using phage display and the yeast two-hybrid screen. Finally, we are developing in vivo metastasis assays using a variety of injection sites to test the ability of KAI1 to suppress multiple cells at various stages of the metastatic process. We are comparing the induction of metastasis of KAI1 positive and negative cell lines following injection at subcutaneous, intravenous and intraperitoneal sites. We are intending to use this data to generate a model that would allow the development of KAI1 gene therapy protocols.